Banks to Put AI Agent Risk on Customers
Wells Fargo's terms take effect November 24, 2026, as other banks restrict third-party agents that log in with customer credentials.
Updated on Nov. 22, 2026 in Artificial Intelligence

Live Poll
Do you trust your bank to protect you if an AI shopping agent makes a costly mistake?
Wells Fargo has enacted an updated customer agreement stating that use of any AI agent is at the customer's own risk, effective Tuesday, November 24, 2026. JPMorgan Chase, Bank of America, Citibank and PNC have also changed terms to restrict or warn against third-party AI agents. These are policy changes, not product launches, and the agents themselves are already in use.
Why it matters
Banks are moving to get ahead of risks from agentic tools, which automate online tasks by clicking on a user's behalf, in case the technology persists as a lasting trend. The policy shift reflects a risk-averse sector treating automated transactions and withdrawals as a security concern.
Wells Fargo's generative AI question-answering tool has been used over a billion times, a figure the source does not place against a time period or baseline. AI agents operate by automating online tasks and clicking on a user's behalf, and the source does not quantify how often such agents have caused errors.
The players
Wells Fargo
A major U.S. bank that operates a generative AI question-answering tool used over a billion times.
JPMorgan Chase
A large U.S. bank that changed its terms in late October to ban third-party AI agents using login credentials.
Bank of America
A major U.S. bank that has warned customers against sharing login information with AI agents and revamped its EricaAssist agent for employees.
Citibank
The consumer banking arm of Citigroup, which has launched an internal development platform centered on agents.
PNC
A U.S. regional bank that changed its terms to include warnings about AI agent use.
The details
AI agents are software programs that carry out online tasks, such as navigating websites or completing purchases, by acting on a user's behalf. JPMorgan Chase's late-October change bans third-party agents that log in using a user's credentials, meaning the username and password used to access an account. Bank of America has warned customers not to share login information with agents, while Citibank and PNC changed their terms to include warnings about agent use.
Timeline
Late October 2026: JPMorgan Chase changes its terms to ban third-party agents that log in with user credentials.
Tuesday, November 24, 2026: Wells Fargo's updated AI agent agreement takes effect.
End of 2026: Multiple banks have enacted new agent policies or warnings.
The Tech Race
Bank of America has revamped its EricaAssist agent to help employees, and Citigroup has launched an internal development platform centered on agents. Those internal efforts run alongside the new restrictions on outside agents, which show the banks deciding where agent access will be permitted.
Customers of the banks named here may find that agents which log in with their credentials are no longer permitted or are explicitly unprotected. Wells Fargo's terms take effect November 24, 2026, and the source does not describe how the other banks will enforce their changes.
The takeaway
Banks are drawing a line between agents that act for customers and agents that log in with their credentials, and the next signal to watch is whether the November 24 Wells Fargo terms are followed by similar changes at other banks. Customer adoption of early agent use is not expected to be significantly deterred by these policy changes, according to Kim.
What happens next
Wells Fargo's updated AI agent agreement takes effect on Tuesday, November 24, 2026, and customers will be bound by its at-your-own-risk terms from that date.
Further reading
For the broader technology context, see our Artificial Intelligence coverage.
Live Poll
Do you trust your bank to protect you if an AI shopping agent makes a costly mistake?





